Research On The Theory Of Hybrid Sliding Mode Control For Complex Electrical Machine Systems And Its Application

As the principle equipments for electrical energy conversion, electrical machine (EM) systems provides a steady strem of strong power for industrial and commercial automation systems. For the diversity of motor types, the complexity of motor's multidimensional and nonlinear mathematical model, the strong coupling among motor current, flux and speed, and the torque disturbances and parameter uncertainties, all these factors result that the classical control theory is unable or difficult to control the EM efficiently. Although the modern control theory has the ability to deal with more complex system, it still cannot fulfill the developing demand on system performance. With the developing of complex industrial drive system, electric vehicle, modern integrated manufracturing system, aerospace technology and multi-links robot, it is important in the sense of academic research and practical value to develop novel control theory and technology with precise tracking and high performance. Sliding Mode Control (SMC) theory is a novel nonlinear analysis and control scheme. Afer nearly 50 year's development, it has grown up as an important branch of modern control theory. SMC control has the merits such as robustness to matched uncertainties and disturbances, fast response ability, simple structure and easy implementation. SMC scheme is researched widely and realized in EM control system successfully. This dissertation integrates the modern control theory and SMC to develop hybrid SMC scheme for complex EM systems. The main contents cover four sections, the adaptive SMC, fuzzy-logic-based SMC, T-S-model-based robust SMC and sliding mode observer (SMO) design.First, for complex EM systems, by combined two kinds of adaptive control, three adaptive sliding mode control schemes are proposed. Scheme 1 is indirect self-tuning SMC for PMSM, and simulation results show that the high-performance speed control with asympotical stability can be obtained in spite of parameter uncertainties and torque disturbances. Scheme 2 is model reference adaptive SMC for induction motor, and simulation results show that it can estimate the rotor resistance and speed onlinely with high accuracy, and also the high performance can be obtained in wide speed range. Scheme 3 is direct self-tuning SMC for a class of uncertain systems. By onlinely tuning the bound layer and control gains, the chattering phenomena is reduced and high-performance tracking is obtained.Second, by utilizing the approximation ability of fuzzy systems, three adaptive SMC scheme are proposed. Scheme 1 is H∞adaptive fuzzy SMC for a class of uncertain systems. This scheme estimates the unknow equivalent control of SMC and H∞technology is adopted to suppress the approximation errors and uncertainties. Scheme 2 is PID-like adaptive SMC for a class of uncertain systems. This scheme estimates the upper bound of uncertainties and disturbances, and PID-like control is adopted instead of switching component of SMC. By estimating the unknown functions with parameter and torque uncertainties, scheme 3 gives indirect adaptive fuzzy SMC for induction motors. Simulation results show that all these schemes can efficiently employ the fuzzy approximation ability to reduce the chattering phenomena with good dynamic tracking and fast convergent rate.Third, by adopted T-S modeling approach, the procedure to model general nonlinear systems are summarized. For a class of uncertain systems with measurable states, the state feeback adaptive fuzzy sliding mode controller is introduced. Then, the T-S approach is employed to model PMSM, and a novel Fuzzy Variable Structure Controller (FVSC) is proposed to implement robust speed control of PMSM. Also, T-S approach is adopted to model induction motor systems, and a H∞fuzzy sliding mode controller (H_FVSC) is then proposed. Simulation or experimental results show that T-S approach has the ability to estabilish proper model of complex motor systems, and all these schemes have good robustness, fast response and good tracking performance.Fourth, three SMOs are proposed to estimate the parameters of EM systems. Scheme 1 is the extended SMO for PMSM, and it can overcome the adverse effect of motor parameters and torque uncertainties on system performance. Scheme 2 is sliding mode current and flux observer for induction motor. It has no integral element and has the merits such as high accuracy, easy implementation and fast convergent rate. Scheme 3 is a novel state-sliding mode observer for induction motor, which is utilized to estimate the motor resistances. Since the SMO is integrated with the state observer, this scheme has better robustness than conventioanl state observer, especially when dealing with large parameter uncertainties. And the number of required parameters is less than that of conventional SMO.At last, the main contents of the dissertation are summarized and the innovations are clearly pointed out. Also, the prospects for future investigation are made.